Slashdot Mirror
Windows 2003 takes 5% away from Linux
ZuperDee writes "According to Netcraft, the number of Windows 2003 servers has doubled since July, and 5% were running Linux before, which is consistent with the trends they've been observing for some time. This doesn't look good for Linux, in my opinion. Maybe we should all start to think about jumping ship?"
Jump ship? Why jump ship? Because others have done so? If I decide to jump off a cliff and fall to my microsoft death will you follow just because you can? Jeez, whats up with people these days.
The article heading is rather misleading. It's not like 5% of all Linux servers converted to Windows Server 2003, or 5% of all servers in the world suddenly run Windows Server 2003. No, of all new Windows Server 2003 installations (which still isn't that many), five percent used to run Linux. It is definitely not time to "think about jumping ship" yet...
/* Steinar */
(This comment is of course GPLed.)
Netcraft now confirms: Linux is dying...
"Perhaps we should all think about jumping ship", eh?
What bollocks. Linux's worth as a server is not judged by its popularity, or its market share. It is, however, judged by how well it performs as a web server, and as a matter of fact it performs very well.
If you're happy and you know it read my blog
And they're replacing the old. No big deal. This is one of the same reasons that Linux got big into the market. Big claims about cost savings, more with less, etc.
If MS new server is a good product, then it should keep the 5% and grow. If it doesn't live up to the hype (replacing 200 servers with 20, saving millions of dollars per annum), its marketshare will dissappear. Initial cost doesn't figure entirely into this either. The software costs for some customers have been subsidised by Bill, and the hardware costs for the upgrades are both minimal, and bugetted because some equipment is becoming EOL'ed by companies three-year plans.
Well, since Sitefinder is running Linux, wouldn't Linux now be running an infinite number more sites than Windows Server 2003?
On the other hand, how many (desktop) Linux converts used to run Windows?
GAAH! MY PRINTER IS ON FIRE!!! PUT IT OUT! PUT IT OUT!
You don't leave a battlefield just because the enemy takes some ground.
Unless it's "Battlefield Earth." In that case just run.
Escape Pod Films: Sketch Comedy and Web Series
What the Microsoft spin doctors do not mention is the continuing market share loss to Apache overall.
When your boss told you to replace your Debian webserver with Win2k3 server, did you...
1[] Capitulate to the microwenie with pleasure.
2[] Change the ID strings and pocket the money.
3[] Install Win2k3 but leave Linux doing the real work (dynamic pages etc).
4[] Tell him to f**k off.
5[] Electrocte the boss with a waffle iron.
So... only 5% chose 1, although some may have done 2 or 3 instead - so call it 2%. Me? My boss knows better than to tell me to install Win2k3 - the previous electrical burn marks attest to that.
Beep beep.
IMHO, if these viruses keep coming around, one is BOUND to attack 2003 servers. Then the 5%'ll feel bad and then revert back.
IMHO as a corporate IT director (and home Linux user), if Linux was:
a) easier to find quality support for
b) able to run more mission critical apps
I would use it in more places in my corporate network. (currently we use it for security and traffic monitoring). I know it is making strong headway, but it is not there yet. I am of course tempted to use it just to spite SCO, but that is not an entirely compelling business reason (the board does not accept your spiteful proposal).
Windows 2003 includes an incredible amount of changes bringing it more into the modular like world of Unix. It will displace some Linux I am sure. I am also sure that Linux will do something which will displace some Win 2K3 that is how this will go for a long while. BTW, understand what you quote about before you make a blatant quote like above. Most of the recent viruses attacked 2K3.
IMHO, and for someone who has been running Windows 2003 server beta on a dev workstation for the last 9 months, I been extremely happy with it.
I have got tons of tools/utils that could bring an XP box to its knees and outright destroy the damn thing. 2003 server has so far been gracefully handling the pressure with no blue screens till last week.
Last week, I came across from first core dump when I was playing around with the Cisco VPN tool and it core dumped (it was due to bad drivers, couldnt find native ones) giving a BAD_POOL_CALLER error a bunch of times.
I thought Xp was way decent than the shitty 98SE and the unbelievably piece of crap ME, but 2003 server has proved that theres a lot of room for improvement. I think they still have a long way to go to capture the server market.
Disclaimer : I have been running a server operating system on a workstation, I admit. Theres guides available to tune the OS to make it run as a workstation and for gaming.
Also, Microsoft has finally shipped an OS with most of its services disabled (including sound) rather than running in to a "gotcha" moment down the line.
Rapid Nirvana
Of all those upgrading to windows 2003, 5% previously used linux. So what?
Compare that to all those upgrading _to_ linux, and look how many of those were previously running other versions of windows? It could easily by a lot more than 5%.
This all looks like a pretty desperate attempt to discredit linux and make win 2003 look more popular than it really is.
Oh, and it's old news anyway.
-- MartinG To mail me: echo kewyjlcxyzvjfxbqwh | tr bcefhjklqvwxyz
Windows 2003 does so badly that it runs only about 0.4% of webservers half a year after release.
Overall IIS loses about 0.2%/month to other webservers.
And now 8500 domains (= 0.002% !) throughout about half a year (= 0.0003%/month) switch from Linux to Windows and people start to get wet their pants.
And then the FUD gets modded as insightful...
Whoah there... I dislike MS quite a bit, but Win2k3 is pretty impressive. For almost a year now I have been testing a beta version in my lab. The box is a lowly PII-200 with 256MB RAM and runs Active Directory, DNS, DHCP, PPTP,and IIS very well. I have been nothing but pleased with it, to my shock and horror. For MS this is a *major* accomplishment.
I don't use Linux because it's an unneccessary pain in the ass to do things with it. I use Win2K Pro. However, all the software running the server components are Open Source (Apache, GuildFTPd) or just well respected freeware like Mercury Mail.
Using Apache just demonstrates what a great product Apache is. It has nothing to do with Linux. I'm not going to abandon the simplicity and stability of Win2K just because Apache can faithfully serve up HTTP requests.
Nobody is debating that IIS is feature bloated hacker friendly piece of garbage. But that has nothing to do with Windows.
I have better things to do with my time (like actually building up the web-site) than dicking around with an OS.
The high quality of one open source product has zero to do with the quality of another.
Ben
Work Safe Porn
...100% of web servers run Apache on Linux, thanks to VeriSign's DNS wildcard being hosted on Apache/Linux.
Why can't I moderate something "Wrong" or at least "Grossly Misinformed"?
> I don't use Linux because it's an unneccessary pain in the ass to do things with it.
Maybe this would be a good time to get specific about what's hard to do on Linux when you're using it for your Web server.
Sheesh, evil *and* a jerk. -- Jade
A piddlingly small percentage of the even more pathetic percentage of sites that chose to try .Not ... er, I mean 2003 Server, we previously using Linux. The meat of the story (such as it is) is that so few sites are even bothering to try 2003 Server.
And I wouldn't be surprised if the story behind the switches from Linux to .Not are mostly cases where a company had their site done by a hosting service (who, sensibly, used Linux) that had grown enough that some twit manager decided they should bring their web presence "in house". Their internal IS people only know Windows, so their obvious choice was 2003 Server (it being perhaps the least bad of the Microsoft stable of shite).
<sigh>
I've had exactly the opposite experience. Once I got rid of all MS servers and moved to Debian, life has been, comparatively, a breeze. I'll take apt-get over hfnetcheck any day! And in comparison, Linux is way, way, way simpler than Windows. Oh, yeah -- and have fun running around to every server in you site to do updates because Windows doesn't support remote admin!
The worst thing that microsoft has done for our industry is to breed a whole generation of check-box programmers and admins -- if they can't tick off some checkbox to perform a task, they don't want to be bothered! What they don't realize is that you can only get 90% of the way there, then you're stuck.
I say this constantly and get modded down for it because I'm not supposed to criticize a "volunteer effort." That attitude right there is problem #1. I don't care if it's a volunteer effort, and neither do most users. We just care about what's sitting in front of us on our screen, the net output.
My point all along has been that people really need to get out of this hobbyist volunteer mindset and realize it's time to create actual results. There's no need to become corporate-minded slaves, but I do wish people would be more professional about things, from project names to interfaces to--and this is the major one--the ridiculous mindset, which you must admit, Slashdot contributes to on a daily basis (usually through "Microsoft hole" articles, when meanwhile my sig shows that distros have more exploits per month anyway...it's all ridiculous).
Professional people admit faults and correct them. We still have some of the same Linux desktop problems as we had five years ago, and people are still complaining about them. Heck, real professional people would zero in on problems before the users even notices them. Professional communities have friendly and courteous tech support, newsgroups, and so on. They have to, because it's all about the customer, i.e., the user. Linux has zealots, trolls, and fanboys. It's not all about the user when it comes to Linux. Mostly, it seems to be about adding enough cool features to be able to take great-looking screenshots for the back of distro boxes, but when you actually grab the mouse to use the thing, it is a disappointing experience (I still remember when GNOME under Red Hat 9 had a stuck taskbar that wouldn't stop moving around with the mouse, and when all else failed and I killed X, of course, that screwed up the boot sequence for some reason...and it was a completely stock install!).
I'm tired of Linux being a hobby OS. Let's face it, outside of the server market (where it is still considered an "alternative OS" despite the fact it has the slight majority), Linux is a hobby OS. The desktop environments are just attempts to SIMULATE a desktop. They don't feel like real, seamless, responsive desktops, but they are written to LOOK like real, responsive desktops, so that people can pretend that they're cool because they use Linux in that way. I wish someone would come out with something so slick and professional that people would have no choice but to switch because of its uber-coolness and usability. This, of course, would call for a complete rewrite, because it would demand things like hardware acceleration, a sane programming API, and so on. I won't hold my breath for it, though. As a matter of fact, the only real uber-cool thing I've seen is Slicker. Its card idea is unique and innovative. Too bad it's tied in with the godawful KDE, but maybe in another few years we'll see things really shine.
But I know that won't happen because people are too busy making yet another toolkit for X or another extension or another weird project with a weird name written all in lower-case on Sourceforge. Meanwhile, in August of 2005, Longhorn is due out, with hardware acceleration, vector-scaled widgets for resolution-independent resizing, a yet-to-be-revealed photorealistic user interface, and even the ability to add and remove RAM without rebooting. I'm sorry, but I don't see all that coming in two years, because two years ago I thought we'd have stuff like that, and two years before that, and so on. It just never comes. And if you request it and wish for it, you get flamed because you're not "doing it yourself." Sometimes it's really easy to despise this community because they refuse to listen unless you're some hero programmer like Linus or Stallman. If you're a user or designer, forget it.
"Sufferin' succotash."
Now I will admit that I still have 2 Linux boxes I host client's web sites on, but most of my boxes now either run FreeBSD or OpenBSD and I have an iBook as my main computer.
Why? It's been my humble opinion that Linux has been a bit of a bastard child. Is it a desktop OS, or a server OS? Its flexablity is its greatest strength and its achellies heel at the same time. With no standards between distros on simple things, like the path to PERL, can cause headaches for software developers. I once work on a project where we had to code three different versions of an app, one for RH, another for Mandrake, and one for Debian. After that expirance, I got fed up with the Linux Platform about the same time as Mac OS X.1 came out.
FreeBSD was/is designed as a server OS first, and if you want to toy with it, it can also make an effective workstation. However this is where Mac OS comes into play: There are companies that are publishing commercial software for the platform. So I can interface wtih 90% of the web design/graphics world that use Photoshop, dreamweaver, QuarkXpress, and other such programs where as due to the pain in the ass Linux is to port across distros, commerical companies WON'T port their products. I will even admit to having MS Office, and I actully LIKE it on mac. It works wonderfully.
While the OSS community has developed some kick ass apps, like the ERP module OSSuite (NOLA I think is the sourceforge project) is what I use to keep track of our business's accounting needs including payroll, W-2's, inventory, etc., there is still a vast void of software needs outthere. GIMP is certianly not a photoshop killer. Back in the days of PS 4 and 5, GIMP looked like it was on the track to possible create a much better product, but as now it seems as though GIMP has made very few improvements over the last two years and it still takes a lot more time and effort to get the same results as Photoshop. Photoshop 7 now blows GIMP away in my book.
The two Linux servers I have still are Sun Cobalt Raq servers and I still use them because of the ease of maintance, but all my ecommerce sites are on FreeBSD machines and I have had very little problems with these boxes. Hell two are still running FreeBSD 3.4 and had uptimes of like 250 days until I patched OpenSSH and several other updates two weeks ago.
RH and SuSE are getting closer to getting Linux from Geekdom to mainstream as SuSE is large in Europe. I used it when I studied over there for semester as the school had a windows lab and a linux lab, but that is mainly a result of GUI installers and KDE & GNOME.
At our new business, I have FreeBSD on both of the terminals (we inheirted two PIII 700 Dells & 3 PIII 550 Gateway's when we bought the business) and instead of paying $2500 for four new computers, I slaped FreeBSD 5 with KDE on there, install mozilla and linked them to the office server which is configured as a local webserver with no outside pipeline and we use OsCommerce as our POS system.
Now this article is a bit trollish about jumping ship. I stats and as Mark Twain wrote, "Lies, Damn Lies, and Stats". Approach with caution. OSS software is starting to get looked at, I work as a independant tech consultant, and Linux gets the press thanks to RH and SuSE providing what Linux needs to get into main stream: commericalization. There is a number you can call for support, if you need it.
"The problem with socialism is eventually you run out of other people's money" - Thatcher.
Linux has achieved tremendous actual results already. Your complaint is that these actual results are not the actual results you're looking for.
Well, I'm sorry, but Linux can't be everything to everybody at all times. I use Linux as my primary desktop and server OS, but unlike you I am not under any delusions that Linux will ever stop being a hobby OS. It is largely written by hobbyists, after all.
This so-called hobby OS of yours still beats windows hands down in areas like multiple virtual desktop support and basic features like including a C compiler. Even the third party virtual desktop managers available for windows (e.g. nvidia deskview, winxp powertoys) have much poorer performance than GNOME and KDE because of the limitations of the windows frame manager API.
That attitude right there is problem #1. I don't care if it's a volunteer effort, and neither do most users.
Frankly, I don't care about your attitude either. Volunteers write software for themselves. They don't write for other people. Let's suppose hypothetically for a moment that the volunteer community were to drop all of their work and concentrate on satisfying your expectations. What tangible benefit would that bring the volunteer community? Answer: nothing. In all likelihood the result would be worse than what we have now, because the motivation is just not there when you're scratching someone else's itch instead of your own.
We just care about what's sitting in front of us on our screen, the net output.
That, my friend, is exactly why volunteers write for their own sake instead of your sake. We're just as selfish as you. We want software that fits our needs, not your needs.
You may try to argue with me on the grounds that Linux somehow "needs" non-developer users like you in order to obtain a sustainable userbase, but what you don't understand is that Linux is not like other commercial operating systems. Because Linux is so volunteer driven, it does not need a large userbase or commercial support in order to thrive in its niche role. The fact that a broader audience might find Linux useful is certainly a nice bonus, but it is not so essential to platform survival that we should sacrifice the core hobbyist nature of Linux to attain it.